This project is currently unmaintained and will remain so for the foreseeable future.
This script and many others like it rely on an unofficial API that has recently become increasingly difficult to support. As Google continues to lock down access to their TTS interface I see no choice other than to suspend maintaining this script for the time being. I sincerely hope that the future will see an official way to use Google TTS on desktop Linux. Until then, please feel free to fork this project if you want to try to fix it.
As a last note: Make sure to also check out the section on similar projects provided in this README.
Ever wanted to use Google text-to-speech on Linux? Now you can.
Table of Contents
The intent of this project is to provide an easy way to use text-to-speech output by Google on your Linux desktop. The script supports reading from standard input, plain text files, and highlighted text. A fall-back interface based on
pico2wave takes care of the TTS output when you are offline.
simple-google-tts is based on
speak.pl by Michal Fapso which uses an unofficial Google TTS API. This results in several limitations.
speak.pl try to work around some of these issues, e.g. requests being limited to 100 characters, but are subject to other restrictions, e.g. obligatory CAPTCHA input for overly frequent requests. Please keep this in mind when using this project.
Why use this script?
Note: What follows is a technical explanation of the inner workings of
simple_google_tts. You don't have to read this to understand how to use this program, but it can help shed some light on the issues you might experience.
Google imposes a 100-character limit on their speech synthesis service that makes it hard to use their TTS system for anything other than short sentences.
speak.pl works around this limitation by breaking the text input down into appropriate chunks. These chunks are set intelligently based on punctuation and syntax of the text. Having processed all chunks
speak.pl then concatenates the speech fragments into one audio file while truncating segments of silence at the start and end of each fragment.
All of these processing steps ensure a relatively natural voice output and minimize the number of clunky pauses caused by the 100-character limitation. The one remaining problem with this approach is that the waiting time between user input and voice playback scales drastically with the length of the text.
This is where
simple_google_tts comes in: Instead of passing the text directly to
simple_google_tts first breaks down the input into paragraphs. The paragraphs are then processed one by one with each paragraph being played back while the next one is synthesized. Any length of text can be parsed with reasonable speed in this manner.
simple_google_tts includes automatic playback, more input modes, an offline TTS back-end, and several adjustments that facilitate parsing of documents with fixed formatting (e.g. selected text in PDF files).
All of this could have probably been accomplished a lot more elegantly within the original
speak.pl script, but I am not familiar with perl.
Installation and dependencies
The following instructions are provided for Debian/Ubuntu based systems.
Overview of all dependencies
You can install all dependencies with the following command:
sudo apt-get install xsel libnotify-bin libttspico0 libttspico-utils libttspico-data libwww-perl libwww-mechanize-perl libhtml-tree-perl sox libsox-fmt-mp3
A breakdown of the dependencies by component and role:
xsel provides support for parsing the X selection contents
libnotify-bin is used for GUI notifications
pico2wave provides the offline speech synthesis back-end
The actual audio playback is handled by
sox, which is part of
Dependencies, as listed in
libwww-perl libwww-mechanize-perl libhtml-tree-perl sox libsox-fmt-mp3
Perl should be part of your default Debian/Ubuntu installation.
Install all dependencies
Clone this repository:
git clone https://github.com/Glutanimate/simple-google-tts.git
Navigate to the download directory
You should be able to run
./simple_google_tts now. If you wish you can symlink
simple_google_tts to your
/usr/local/bin) to make it easier to access.
speak.pl must always reside in the same directory as
simple_google_tts <options> <languagecode> <input>
$ simple_google_tts en "Hello World" Reading from string. Using Google for TTS synthesis. Synthesizing virtual speech. Processing 1 out of 1 paragraphs Playing synthesized speech 1 All sections processed. Waiting for playback to finish.
simple_google_tts can read text from standard input or a text file. The syntax is the same in each case. The script will automatically identify the type of input provided and perform the text to speech synthesis via
If no arguments are provided
simple_google_tts will try to read from the current X selection. This corresponds with the currently highlighted text. Using this functionality you can set up a keyboard shortcut that automatically reads out selected text.
At all times you can access an overview of all supported options by invoking the help output:
$ simple_google_tts -h simple_google_tts [-p|-g|-h] languagecode ['strings'|'file.txt'] -p: use offline TTS (pico2wave) instead of Google's TTS system -g: activate gui notifications (via notify-send) -h: display this help section Selection of valid language codes: en, es, de... Check speak.pl for a list of all valid codes Warning: offline TTS only supports en, de, es, fr, it If an instance of the script is already running it will be terminated. If you don't provide an input string or input file, simple_google_tts will read from the X selection (current/last highlighted text)
-p: By default
speak.plto query Google's speech synthesis service and only fall back to
pico2waveif no Internet connection is found. If you don't want to use Google's TTS service you can use this option to default to
-g: If you plan to assign
simple_google_ttsto a keyboard shortcut you can use this option to enable GUI notifications using
libnotify-bin(the default notification daemon).
-h: Display help section
Google's TTS service currently supports the following language codes:
af Afrikaans sq Albanian am Amharic ar Arabic hy Armenian az Azerbaijani eu Basque be Belarusian bn Bengali bh Bihari bs Bosnian br Breton bg Bulgarian km Cambodian ca Catalan zh-CN Chinese (Simplified) zh-TW Chinese (Traditional) co Corsican hr Croatian cs Czech da Danish nl Dutch en English eo Esperanto et Estonian fo Faroese tl Filipino fi Finnish fr French fy Frisian gl Galician ka Georgian de German el Greek gn Guarani gu Gujarati ha Hausa iw Hebrew hi Hindi hu Hungarian is Icelandic id Indonesian ia Interlingua ga Irish it Italian ja Japanese jw Javanese kn Kannada kk Kazakh rw Kinyarwanda rn Kirundi ko Korean ku Kurdish ky Kyrgyz lo Laothian la Latin lv Latvian ln Lingala lt Lithuanian mk Macedonian mg Malagasy ms Malay ml Malayalam mt Maltese mi Maori mr Marathi mo Moldavian mn Mongolian sr-ME Montenegrin ne Nepali no Norwegian nn Norwegian (Nynorsk) oc Occitan or Oriya om Oromo ps Pashto fa Persian pl Polish pt-BR Portuguese (Brazil) pt-PT Portuguese (Portugal) pa Punjabi qu Quechua ro Romanian rm Romansh ru Russian gd Scots Gaelic sr Serbian sh Serbo-Croatian st Sesotho sn Shona sd Sindhi si Sinhalese sk Slovak sl Slovenian so Somali es Spanish su Sundanese sw Swahili sv Swedish tg Tajik ta Tamil tt Tatar te Telugu th Thai ti Tigrinya to Tonga tr Turkish tk Turkmen tw Twi ug Uighur uk Ukrainian ur Urdu uz Uzbek vi Vietnamese cy Welsh xh Xhosa yi Yiddish yo Yoruba zu Zulu
Please note that, out of these, the
pico2wave back-end only supports the following languages:
en English de German es Spanish fr French it Italian
Read English text from file
simple_google_tts en readme.md
Read from X selection, using pico2wave, and enable notifications
simple_google_tts -gp en
to prevent simultaneous output the script tries to force only one instance at a time. Unfortunately this fails sometimes, which can be a problem when using the script through a keyboard shortcut
there is no easy way to terminate the TTS output if the script is used via a keyboard shortcut.
Highlighting an empty line or space and then executing the script should, in theory, terminate the last script instance and stop the playback. Because of the first issue this does not always work.
You could probably assign another hotkey to terminate any running instances of the script (e.g.
pkill -9 simple_google_tts; warning: I have yet to try this out).
too many requests too quickly will cause Google to start requesting CAPTCHA input. I have yet to hit this limit in my regular use of the script.
speak.pl copyright 2012 Michal Fapso
simple_google_tts copyright 2014 Glutanimate
This project is not endorsed, certified or otherwise approved in any way by Google™.